Endurance

Endurance was a wooden three-masted barquentine with a length of 43.9 m, beam of 7.6 m and draught of 3.5 m. The ship had also a 350 hp steam engine. Endurance was designed by Ole Aanderud Larsen and built at the Framnæs shipyard in Sandefjord, Norway, where Johan Jacobsen was responsible for the construction. The original name of the vessel was Polaris, given by Adrien de Gerlache and Lars Christensen, who were the first owners together with other investors. Construction of Polaris started in early 1911, and the ship was launched in December 1912 (Verlinden, Reference Verlinden2017; Mitchener, Reference Mitchener2015, p. 190).

The structure of Polaris is shown in Figures 1–3. The ship had three decks: Lower (tween) deck, main deck and, above those, a short bridge deck. Only the main deck covered the whole ship’s length. The lower deck was interrupted at the machine room to make space for the steam engine and boiler. In Figure 1, the machine room can be located by the funnel and boiler, and in Figure 3, by the large open area on the lower deck drawing. In the machine room area, only one deck beam spanned the whole breadth of the ship, and five deck beams were discontinuous. The machine room area was, therefore, a weak part of the ship.

Figure 1. Endurance. Cross-section along the centre line, showing the hull structure (Framnæs Mekaniske Værksted, 1911). Reproduced by kind permission of the Vestfoldmuseene.

Figure 2. Midship cross-section of Endurance (Framnæs Mekaniske Værksted, 1911). Reproduced by kind permission of the Vestfoldmuseene.

Figure 3. Decks of Endurance. From the top: bridge deck, main deck, and lower (tween) deck (Framnæs Mekaniske Værksted, 1911). Reproduced by kind permission of the Vestfoldmuseene.

The keel was made of oak, four members on top of each other, making a total height of almost 1.5 m (Fig. 2). About 300 mm of the keel was below the hull planking. Frames were oak near the bottom and pine higher up. Planking was pine, sheathed with greenheart. Deck beams were also made of pine, and the connections of beams and frames were strengthened with knees; made of spruce for the main deck and of iron for the lower deck. The bow was strengthened with an iron rail and iron plates.

Polaris was designed and built for polar tourism, for hunting polar bears and walruses in the Arctic. Maybe due to the First World War looming, maybe because de Gerlache and Christense ran out of money, Polaris did not make a single voyage to the Arctic and in March 1914 was sold to Shackleton, who changed its name to Endurance (Verlinden, Reference Verlinden2017).

Weddell Sea 1915

The events that led to the loss of Endurance in the Weddell Sea in 1915 are told below using the diaries and memoirs written by members of the expedition. This paper draws on the diaries of Frank Worsley (Reference Worsley1916), Harry McNish (Reference McNish1916), Reginald James (Reference James1916), James Wordie (Smith, Reference Smith2004), Frank Hurley (Dixon & Lee, Reference Dixon and Lee2011), Thomas Orde-Lees (Thompson, Reference Thompson2020), and Ernest Shackleton (Reference Shackleton1915), together with the memoirs of Worsley (Reference Worsley1931) and Shackleton (Reference Shackleton1920). There are some differences in the stories, but no contradictions. Memoirs were also written by Frank Hurley (Reference Hurley1925), Leonard Hussey (Reference Hussey1949) and Frank Wild (Butler, Reference Butler2011), but they do not add any relevant details to the account.

Endurance became beset in sea ice in the Weddell Sea on 18 January 1915. A few attempts were made to free the ship, but on 24 February, watches were ceased and the ship became a floating research station drifting northwards with the ice (Worsley, 24.2.1915). Between February and October, when Endurance was finally crushed and abandoned, there were five serious events that could have sunk the ship, but the events ended before the forces involved grew too large.

The first serious event occurred on 4 April. Wordie, one of the scientists, heard loud rumbling noises: there was pressure in the ice and an ice ridge, about three metres high, formed near the ship. At the same time, Endurance was vibrating slightly. (Worsley, 4.4.1915; Wordie, 3.4.1915. Wordie’s diary is occasionally out by one day, sometimes earlier and sometimes later than the other diaries.) If the sail, the above-water part of a ridge, is three metres high, the underwater part of the ridge reaches to a depth of more than ten metres. The vibration, in turn, suggests that ice was failing through a compressive failure process, crushing directly against the ship hull. Crushing of ice results in high ice loads on a small area, high enough to break ship structures.

The second noteworthy event started on 14 July when a violent snowstorm hit Endurance, and noises resembling the breaking of ice were heard below the ship. Ship’s carpenter McNish rushed on deck but could not figure out where the noise was coming from. According to Wordie, Shackleton was very nervous on that day and went several times on deck to check the situation. Boxes stored on deck were carried below, and the deck was cleaned of snow. Shackleton asked McNish to attach runners to a lifeboat. Over the next few days, pressure ridges kept forming in the one metre-thick ice around Endurance, and Shackleton, together with his second in command Frank Wild and ship’s captain Frank Worsley, started four-hour watches on deck. Food, matches, oil, and other important items were collected on deck. McNish wrote in his diary on 24 July: “We have everything prepared for leaving the ship”. (McNish 14.–24.7.1915; Wordie 14.7.1915)

The third dangerous event took place on 1 August. Endurance made an abrupt movement, and grinding sounds were heard under the ship. The expedition dogs were at once moved from the ice to the ship, and soon after, the ice floe fractured and ridges started to form near the ship. The ice piled under the ship, lifted it up and heeled first to starboard (right) and then to port. Deck beams buckled, and some of the trenails (wooden pegs) got started under the pressure from the moving ice. Shackleton encouraged everybody to rest as much as possible and to keep warm clothes either on or nearby. Captain Worsley packed his chronometers, sextants and tables, the scientists packed their notebooks, and carpenter McNish placed photos of his family between pages of a Bible, and the Bible into his bag. (James 1.8.1915; Worsley 1.8.1915; Wordie 1.8.1915; McNish 1.8.1915)

Compressive events came and went, and gradually the expedition members got used to them, until 30 September, when the situation escalated fast. Ice started crushing against the ship’s hull, which experienced vibrations violent enough to shake the whole rigging and make boxes fall down from shelves. Frames were bent inwards, the tween deck was forced up, bulkheads inside the ship bulged, and linoleum on the floors buckled. It looked like Endurance “was going into pieces”, as McNish put it, but then the loading ended. This was the fourth of the serious events. (James 30.9.1915; Orde-Lees 30.9.1915; McNish 30.9.1915)

The fifth dangerous ice loading event started on 17 October. Again, the ice piled under Endurance. The vessel rose up about a metre and heeled to port, it was possible to see about 20 cm of the propeller. At the same time, loud screeching noises were heard from the engine room where the iron plates on the floor buckled up and rafted over one another. Reginald James, one of the scientists, wrote that “for a time things were not good as the pressure was mostly along the region of the engine room where there are no beams of any strength” (James 17.10.1915). Captain Worsley noted on the same day that “engine room [was] the weakest part of the ship” (Worsley 18.10.1915). Both were right. In the afternoon the following day, events escalated. Endurance was still heeled to port and rested against a floe on the port side. At 4.45 p.m., the ice floe broke up, and Endurance heeled over to almost 30 degrees to port. At that stage, the keel was visible from the starboard side. Worsley admitted thinking that it was “possible for the floe to force us right over & even override us” (Worsley 18.10.1915). But at 7.45 p.m., the ice around the ship started to fracture and gradually Endurance come upright again. (James 17.10.1915; Worsley 17. & 18.10.1915)

The end of Endurance started in earnest after dinner on 24 October. The crew heard a loud crash, which was followed by a prolonged vibration of the whole ship. A floe coming from the starboard side had hit the stern and twisted the rudder and the sternpost, with the result that planks parted from the sternpost on the starboard side. The ship started to leak. However, McNish was able to build a cofferdam in the aft section of the machine room and, together with pumps, it was fairly easy to keep the leak under control. (Worsley 24.10.1915, James 24.10.1915, Hurley 24.10.1915)

Worsley’s diary from that day has a drawing of Endurance nipped by three ice floes, with arrows showing the motion of the ice. The drawing illustrates how the bow is held by one floe, another floe is loading the stern from the starboard side, while a third floe on the port side is loading Endurance at midships (Fig. 4). A drawing of the same event in Shackleton’s book South (1920, p. 72) is similar, but a mirror image, with the stern loaded by an ice floe from the port side. The underwater images of the wreck of Endurance show the rudder at the sea bottom on the port side of the ship (Morelle & Francis, Reference Morelle and Francis2024; Addley, Reference Addley2024), suggesting that the drawing in Worsley’s diary is correct.

Figure 4. Endurance nipped by ice on 24 October 1915 (Worsley, Reference Worsley1916). Reproduced by kind permission of the University of Cambridge, Scott Polar Research Institute.

Worsley’s diary has other important information from that day. Together with officers Lionel Greenstreet and Hubert Hudson, Worsley went down to the hold of Endurance to clear the ice and dirt from the bilges; the pumps were not working as the pipes were frozen up. The men succeeded in their task, but the work in the cold and darkness was terrifying, and they were relieved when they got back in the open air. A part of the unpleasant experience was the noise within the ship’s hull. In his diary Worsley wrote that “…this is not a pleasant job…with the beams & timbers groaning & cracking all round us like pistol shots”, and later, in his memoirs say that “we continually expected the ship was going to collapse on top of us and bury us alive” (Worsley, 24.10.1915; Worsley, Reference Worsley1931, p 15). From this, it is evident that, at this stage, Endurance was breaking up below the tween deck at midships; not only at the stern near the waterline, where the sternpost and planks met and where the leak was.

The next day, 25 October, was uneventful, but on 26 October, the ice pressure returned. Pressure ridges were forming near Endurance, and gradually the active ice failure zone reached the ship. Endurance was bent like a bow, the bulwark at midships broke in, windows splintered, deck beams bent, ends of planking on the starboard side opened more than 10 cm, the sides of the ship were bent inwards, and masts were shaking as the tension in the stays varied with the deforming ship. Ice pushed under the ship and lifted it up, which reduced the leak somewhat. (Worsley 25.10.1915; James 25.10.1915; Hurley 25.10.1915; Shackleton, Reference Shackleton1920, p. 71–72)

Wednesday, 27 October 1915, was a sunny day on the Weddell Sea. There was pressure in the almost 1.5-metre-thick ice all day, but at 4 p.m., the pressure intensified, raising first the bow and then the stern more than a metre out of the water. Endurance was heavily listed to starboard, and a large part of the bottom was exposed from the port side. “The rudder and propeller were buried in a maze of pressure blocks”, as James Wordie put it (Wordie 27.10.1915). Endurance had become a part of a pressure ridge, and rubble was up to the bulwark level on the starboard side. Then the rudder, sternpost, and part of the keel were torn off; deck beams and decks broke up; the side planking fractured. As a result of all this damage, Endurance was filling rapidly with water, and Shackleton ordered the crew to abandon ship, and they got off onto a nearby ice floe. (Worsley 27.10.1915; McNish 27.10.1915; James 27.10.1915; Hurley 27.10.1915; Wordie 27.10.1915; Shackleton, Reference Shackleton1915; Shackleton, Reference Shackleton1920, p. 76; Worsley, Reference Worsley1931, 17–19)

The expedition abandoned Endurance on 27 October, but it was not until 21 November that the vessel sank (Shackleton, Reference Shackleton1915). During those weeks, Endurance was full of water, but was held on the surface by the buoyancy of the ice rubble under the ship and by the frictional forces maintained by the compressive ice. This was well understood and described by carpenter McNish: “she is down level with the decks now but there is a lot of ice under her & when the floe opens we expect she will sink as she is broken in two halfs” (McNish 9.11.1915). The last six words in his note are important.

The structural damage to Endurance can be categorised into four groups, according to their location on the ship:

1. 1. The rudder and the sternpost. They were torn off, which is mentioned in all the diaries and is also confirmed by the underwater images from the wreck site (Morelle & Francis, Reference Morelle and Francis2024; Addley, Reference Addley2024). The rudder is at the bottom, on the port side of the stern, in one piece. Beside the rudder is a partly buried beam, which could be the sternpost, based on the location and dimensions as compared to the rudder.

2. 2. The keel. McNish wrote in his diary on 27 October that “the keel was ripped off” (McNish 27.10.1915). His observation is mentioned on the same day by Wordie, who shared a tent with the carpenter: “Chippy tells me from his bag that the keel went during the pressure” (Wordie 27.10.1915). Captain Worsley made the same observation as McNish and wrote later in his book: “… part of the keel was driven upwards by the ice” (Worsley, Reference Worsley1931, p. 17). As the ship was heavily listed to starboard, driving the keel upwards presumably means driving the keel along the ship’s bottom on the port side. The note by McNish, that Endurance was broken into two halves (McNish 9.11.1915), is a logical consequence of the failure of the keel. The underwater images show, at some distance forward from the rudder, on the port side, something that could be the end of a displaced part of the keel sticking up from under the ship (Morelle & Francis, Reference Morelle and Francis2024; Addley, Reference Addley2024). The dimensions are similar to those of the beam that was suggested above to be the sternpost. The location on the port side is logical considering the chain of events described above.

3. 3. Deck beams. Buckling and breaking of deck beams are mentioned in all the diaries. This is due to compressive loads on the hull and is well described by Shackleton: “Just before leaving [the ship], I looked down the engine-room skylight as I stood on the quivering deck, and saw the engines dropping sideways as the stays and bed-plates gave way” (Shackleton, Reference Shackleton1920, p. 76). The engine room was a large open space without beams and, as Captain Worsley acknowledged, the weakest part of the ship (Worsley 17.10.1915).

4. 4. Ship’s sides. “The ice has cut clean through the ship”, wrote McNish in his diary on 28 October (McNish 28.10.1915). The same was reported by Worsley: “Great spikes of ice were now forcing their way through the ship’s sides” (Worsley, Reference Worsley1931, p. 18–19). The underwater images of the wreck of Endurance do not show great holes in the sides, but those mentioned by McNish and Worsley may well be buried in the mud on the seabed. On 27 October, Endurance was heavily listed and the bottom was exposed. There may be damage there, but we may never know.

The idea of the rudder as the Achilles’ heel, as the sole or main reason for the loss of Endurance, is not supported by the diaries or memoirs of the expedition members. They had seen their ship demolished by the pressure of the Weddell Sea ice and had described it in detail. What happened to Endurance on 27 October 1915 can be summarised by the words of Shackleton: “I cannot describe the impression of relentless destruction that was forced upon me as I looked … around. The floes … were simply annihilating the ship” (Shackleton, Reference Shackleton1920, p. 76).

What did Shackleton know before the expedition?

What did Shackleton think of the strength of Endurance and of its suitability for the Weddell Sea before the expedition? In his book South, Shackleton wrote that in building Endurance, “shipwrights had never done sounder and better work”, and continued: “… no ship ever built by man could live if taken fairly in the grip of the floes and prevented from rising to the surface of the grinding ice” (Shackleton, Reference Shackleton1920, p. 66 & 62), giving support to the narrative of an exceptionally strong ship that faced an insurmountable natural force, the Weddell Sea ice. However, as will be explained below, Shackleton was fully aware of the weaknesses of Endurance and the risks involved when sailing into the Antarctic pack ice.

Antarctic, Scotia, and Deutschland

Endurance was the fourth research vessel to sail into the Weddell Sea. Antarctic, the ship of the Swedish expedition led by Otto Nordenskjöld, was caught in compressive ice in November 1902 and sank in February 1903. The story of Antarctic is very similar to that of Endurance. After Antarctic sank, the men moved onto an ice floe, sailed with a lifeboat to a remote island, built a hut from rocks and the lifeboat, and in the end, everybody was saved. The two ships were also crushed by ice in the same way: they were lifted up by the compressive ice, heeled over, vibrating due to the ice crushing on the ship sides; beams buckled, planking broke, rudder and sternpost were damaged and, finally, the end came when the keel was ripped off by an advancing ice floe sliding along the bottom and hitting it. All this is vividly described in the book written by the Swedish expedition (Nordenskjöld et al., Reference Nordenskjöld, Andersson, Larsson and Skottsberg1904). The book also has two interesting notes (Nordenskjöld et al., Reference Nordenskjöld, Andersson, Larsson and Skottsberg1904, Part II, p. 423): “All ships designed for the Arctic Ocean would presumably get damaged in Antarctic waters”, and “In short, if Antarctic did not have such a [protruding] keel, she would still be sailing the oceans”.

Shackleton knew the fate of Antarctic very well. He had been involved in the rescue operation, and the book of the Swedish expedition, translated into English, was in the library on board Endurance. Shackleton even took the book with him when Endurance sank (Shackleton, Reference Shackleton1920, p. 229).

In February 1903, at the same time as the ice was crushing the Antarctic, a Scottish expedition led by William Speirs Bruce was trying to reach the Antarctic continent on the eastern side of the Weddell Sea. Bruce was careful not to let his ship, Scotia, get nipped by ice, and returned north for winter. The following year, ice conditions were easier, and in March 1904, Scotia reached the ice shelf. This time, Scotia could not escape compressive ice and ice blocks piled against its sides, some ending on the deck. The ice piled up also under the ship, and it was lifted up about a metre. When the compressive situation ended and the ship was freed, Bruce sailed north and escaped from the ice. (Day, Reference Day2013, p. 122; Bryan, Reference Bryan2011, p. 190)

Deutschland was the third expedition ship that sailed to the Weddell Sea. Originally called Bjørn, the ship was built in 1905 for whaling and sealing in the Arctic. Shackleton had considered buying Bjørn for his 1907–1909 Antarctic expedition, but did not have enough funds and ended up buying the cheaper Nimrod (Riffenburgh, Reference Riffenburgh2004, pp. 121–123). In 1910, Wilhelm Filchner bought Bjørn for the second German South Polar Expedition, renamed the ship Deutschland, and took it for modifications to the Framnæs shipyard in Sandefjord, Norway. From there, in February 1911, Deutschland sailed to Hamburg, Germany, for final modifications before going south (Filchner, Reference Filchner1922).

Shackleton helped Filchner in his preparations for the expedition (Mill, Reference Mill1923, p. 185). In early 1911, Shackleton visited Sandefjord while Deutschland was there and advised Filchner to strengthen the ship’s hull for the Weddell Sea ice by adding diagonal beams to the ship (Filchner, Reference Filchner1922, p. 26). Filchner followed that advice; the modified hull structure of Deutschland is shown in Figure 5.

Figure 5. Deutschland. Midship cross-section after modifications suggested by Shackleton (Filchner, Reference Filchner1922/1994). Reproduced by kind permission of the Erskine Press.

Another important event took place while Deutschland was in the Framnæs shipyard. In January 1911, the shipyard signed a contract to build Polaris, which would later become Endurance (Verlinden, Reference Verlinden2017).

Filchner and his expedition reached the southern extremity of the Weddell Sea, the ice shelf that is now called the Filchner-Ronne ice shelf, in early 1912. In March, the expedition headed north, but Deutschland got beset in ice. The ice crushed against the hull, and the ship vibrated the same way as Antarctic had ten years earlier and Endurance would three years later. For eight months, Deutschland drifted with the ice, but finally broke free and sailed home.

The crucial difference between Antarctic and Endurance, which were crushed by the Weddell Sea ice and sank, and Deutschland, which survived the compressive ice, were the diagonal supports added in Sandefjord into Deutschland. It is noteworthy that this strengthening was planned in accordance with advice from Shackleton and was carried out by the Framnæs shipyard, which was in the process of building Endurance. Therefore, both Shackleton and the shipyard knew how to take compressive sea ice into account in ship design, and after Deutschland returned from the Weddell Sea, both knew that the design had been successful. Why then the hull of Endurance was not strengthened in a similar way, after Shackleton had bought the ship in March 1914, is not known.

Ships in compressive ice – a comparison

Endurance was not built for Weddell Sea pack ice conditions, or for polar winter at all. Endurance was built to take tourists to the Greenland Sea, Svalbard, Novaya Zemlya and Jan Mayen during the Arctic summer, from June to September (Verlinden, Reference Verlinden2017). The ice conditions at the edge of the polar pack in the Arctic are very different to those deep inside the pack ice in the Antarctic. At the ice edge, ships are mostly dealing with collisions with ice floes. Endurance was designed for this kind of ice conditions, and its planking and frames were thick and strong enough for that task. However, in pack ice conditions, where compression from the ice needs to be taken into account, deck beams become of key importance. It is the deck beams that keep the two ship sides apart and maintain the shape of a ship. Without strong enough deck beams, a vessel gets crushed by compressive ice, more-or-less irrespective of the thickness of planking and frames. This is where Endurance was not on a par with other early polar ships. There were also other design details that challenge the view that Endurance was a particularly strong ship.

Table 1 shows a list of early Antarctic ships. As mentioned above, Scotia was used by William Speirs Bruce in his 1902–1904 expedition. Scotia was built as a whaler, and before sailing south, it went through significant modifications planned by G.L. Watson. Aurora was also built as a whaler and was refitted for the 1911–1914 Australasian Antarctic Expedition led by Douglas Mawson. The refitting was supervised by John King Davis, who became the captain of the ship.

Table 1. Main dimensions of early Antarctic ships. L is the length, B is the breadth, L _M is the length of machine room area with discontinuous tween deck, and N is the number of stanchions supporting deck beams, see Figure 6 (Bryan, Reference Bryan2011; Mitchener, Reference Mitchener2015). For Aurora, N is not known. The midship cross-section of Belgica (de Gerlache, Reference de Gerlache1938) does not show a single stanchion, but other drawings of the ship suggest the decks had some support.

The first ship to winter in Antarctica, Belgica, was built in 1884 and used in whaling before being purchased and refitted by Adrien de Gerlache for the 1897–1899 Belgian expedition. Belgica has interesting connections to Endurance. Belgica was designed and constructed in Selvik, Norway and refitted for the expedition in Framnæs shipyard in Sandefjord. In both shipyards, the work was led by Johan Jacobsen. About ten years later, de Gerlache designed Polaris, which was built in Sandefjord under the leadership of Jacobsen and later sold to Shackleton, as mentioned above. According to Kjær (Reference Kjær2005), “the design of Polaris was based on de Gerlache’s experience with Belgica”. This is easy to understand by looking at the drawings of the two ships, they are very similar in construction. Maybe de Gerlache wanted Polaris to be an updated, luxury version of Belgica.

Fram, Gauss and Discovery were all designed and built for polar exploration. Discovery was built for the British 1901–1904 expedition led by Robert Falcon Scott. Although a purpose-built expedition ship, the design of Discovery followed more-or-less the tradition of whalers and sealers. Fram and Gauss were altogether a different kind of polar ships.

Fram was designed by Fridtjof Nansen and Colin Archer for pack ice, to survive when nipped in compressive ice. Fram was first used by Nansen in his 1893–1896 attempt to reach the North Pole by drifting with the ice, and then by Roald Amundsen in his 1910–1912 Antarctic expedition. In the construction of Fram, Nansen and Archer understood two important features: (1) The shape and size of the ship should be such that it will rise when the ice presses against its sides; (2) The vessel should be made strong enough to withstand the greatest possible pressure from ice, from any direction whatsoever (Nansen, Reference Nansen1897, p. 60). Fram was successful and showed that a wooden ship can be designed for compressive ice. The design was then followed in construction of Gauss, used by Erich von Drygalski in his 1901–1903 Antarctic expedition (Drygalski, Reference Drygalski1904).

Deck beams

Figure 6 shows two idealised cross-sections of early Antarctic ships under compressive loading from ice. The deck beams carry the horizontal load and are themselves supported by vertical supports, stanchions. For a beam in compression, the height (h) and width (b) of the beam cross-section, as well as the span (ℓ), are important. The compressive force F _B that makes a beam buckle is related to h, b and ℓ as follows:

(1) $${F_B}\;\sim\;{{{b{h^3}}}\over{{{\ell ^2}}}}$$

Figure 6. Idealised cross-sections of early Antarctic ships. Thick arrows represent compressive ice load. Endurance, Discovery, Belgica, Scotia, and presumably Aurora, were of the type (a); Fram, Gauss and Deutschland were of the type (b).

Scotia, Discovery, and Endurance were of the design shown in Figure 6(a), and the dimensions of their deck beams are compared in Table 2. The midship cross-sections of these ships show one stanchion at midships and thus ℓ ≈ B/2, where B is the ship breadth. In addition to their dimensions, the number of the deck beams also matters. Table 2 shows the number of beams per metre (N/L). By using N/L × bh ³/ℓ² as a proxy for the compressive load a ship can sustain, it can be calculated that Scotia carried about 2.7 times higher load than Endurance. Discovery, with higher ℓ, was about 10% weaker than Endurance in this respect. The midship cross-section of Belgica (de Gerlache, Reference de Gerlache1938) does not show a single stanchion, but other drawings of the ship suggest that at least some of the deck beams were supported at midships. By using ℓ ≈ B/2 for Belgica also, it can be estimated that Belgica carried about 1.7 times higher load than Endurance. However, if the deck beams of Belgica were unsupported, their strength would have been less than half of the strength of deck beams in Endurance. Presumably, Aurora also had the kind of hull shown in Figure 6(a), but drawings showing its midship cross-section were not found. It is assumed in these comparisons that the material properties of the wood used in the different ships were similar.

Table 2. Dimensions of deck beams (main deck) and frames in Scotia, Belgica, Discovery, and Endurance. h is the beam/frame height, b is the beam width, ℓ is the beam span (distance between supports), N is the number of beams on the ship, and L is the ship length (Bryan, Reference Bryan2011; de Gerlache, Reference de Gerlache1938; Framnæs Mekaniske Værksted, 1911). f _DB and f _FR indicate the relative ability to carry horizontal loads for deck beams and frames, respectively, normalised to give unity to Endurance; f _DB ∼ N/L × bh ³/ℓ², f _FR ∼ h ².

Fram, Gauss and Deutschland were of the design illustrated in Figure 6(b). These ships had three stanchions and diagonal supportive beams. Both the additional stanchions and the diagonal supports are very effective in stiffening and strengthening a ship hull.

The importance of the stanchions comes through the support they give to the deck beams. The modifications suggested by Shackleton to Deutschland decreased the deck beam span (ℓ) and more than doubled the strength under compressive loading. By using 1/ℓ² to compare ships with different designs, it can be calculated that, due to shorter deck beam spans only, Fram and Gauss would have carried about two times, and Deutschland more than one and half times larger forces than Endurance. This comparison considers the effect of deck beam span only and ignores all other dimensions. In addition to shorter beam spans, Fram and Gauss had more than 1.4 beams per metre, compared to 1.05 in Endurance.

The importance of the diagonal supports comes through their ability to resist distortion of a ship hull. A compressive loading is symmetric only when the forces act on both sides at the same height, and a ship is not heeled. However, an ice load becomes easily non-symmetric: it is enough that the ice piles against a ship hull on one side in a different way than on the other side. Such a non-symmetric compressive loading makes a ship heel, but more importantly for this analysis, it distorts the hull. When this happens, the diagonal supports in Figure 6(b) become very effective and make a ship significantly stiffer and stronger than a ship without any diagonal supports.

Engine room

Due to the size of the engine and the boiler, many early polar ships, Endurance included, lacked a tween deck in the engine room. This created a large open space which, in the longitudinal direction, was defined by a bulkhead in front of the engine room and another behind it. In the vertical direction, the open space was defined by the main deck and the keel. From a structural point of view, the missing tween deck is a challenge, as illustrated in Figure 7. First, with the tween deck missing, all the compressive loading is carried by the main deck, and the strength of the hull is reduced by roughly half. Second, it is difficult to arrange stanchions in an engine room. This increases the span of the deck beams, and thus decreases their strength. Third, the frames lack all the support in the area limited by the main deck and the keel in the vertical direction. The larger the open engine room is, the weaker a ship is.

Figure 7. Idealised ship cross-section from a machine room area. Note the lacking tween deck. Compared with the cross-sections illustrated in Figure 6, both the deck span and the frame spans have increased.

Table 1 gives the length of the machine room for the ships studied. At 8.4 m, 22% of the ship’s length, the machine room of Endurance is the longest of all the early Antarctic ships, both in metres and in percentage of the ship length. Figures 1 and 3 show the structure of Endurance in the machine room area. In addition to the missing tween deck, the main deck has only one continuous deck beam for the whole 8.4 m machine room length. In Fram and Gauss, the machine room problem is avoided by squeezing the engine and boiler between the decks, allowing the tween decks to extend the whole ship length.